6.3.4 Resonance and Laser-Induced Fluorescence  6.3 Measurements of Chemical Composition  6.3.2 Remote Sensing from the Ground and Aircraft

6.3.3 Tunable Diode Laser Absorption Spectroscopy

Up until this point, remote sampling spectroscopic measurements of trace atmospheric gases have been described. Such techniques employ long path sampling through the open atmosphere where the sample is not drawn through a sampling system. Although this technique avoids sampling problems, the interrogated volume is better defined using in situ techniques, where the ambient sample is drawn through an instrument or into a grab canister. Measurement techniques employing in situ sampling will be covered in the remainder of this chapter. The first such technique to be discussed is tunable diode laser absorption spectroscopy.

Tunable diode laser absorption spectroscopy (TDLAS) is a highly selective and versatile technique for measuring many trace atmospheric constituents with detection sensitivities in the sub-parts-per-billion (ppbv) concentration range. As the name implies, this technique utilizes a tunable diode laser source which emits in the mid-infrared spectral region between 3 and 30 $\mu$m. Individual lasers, which are comprised of tertiary or quaternary salts of lead, are tailor made to access specific regions of the mid-infrared spanning tens to several hundred cm^-1 in width. The lasers are cooled cryogenically to a temperature between 10-120 K. By adjusting the laser conditions, generally the temperature and/or injection current, the output wavelength of individual devices can be tuned continuously in small intervals of several cm^-1 throughout the entire tuning width. Although each continuous interval is separated by several cm^-1, one can sometimes gain access to the gaps by further adjustments in laser temperature and current.

The mid-infrared spectral region accessed by diode lasers is extremely attractive for detecting many atmospheric species with high selectivity. Many trace gases of atmospheric importance exhibit moderate to strong absorptions in this region while the major constituents, oxygen and nitrogen, do not. Furthermore, absorption lines in this region, which result from vibrational-rotational transitions, predominantly appear as sharp discrete features for small molecules when the sampling pressure is in the 1 to 50 torr range. Under such conditions, the spacing between individual absorption features generally exceeds typical absorption linewidths, which for most atmospheric molecules are in the 0.001 cm^-1 range. Tunable diode laser (TDL) linewidths by contrast, are typically in the 10^-4 to 10^-5 cm^-1 range. These conditions result in the high resolution, and hence high selectivity, inherent in TDLAS. This is in contrast to spectroscopic measurements in the visible and ultraviolet regions where many atmospheric species exhibit broad and non-structured absorption features which can overlap.

Figure 6.5 schematically shows a general setup which is fundamental to all lead-salt tunable diode laser systems:

Figure 6.5:  A schematic representation showing a general setup which is fundamental to all tunable diode laser systems.

The sample of interest is drawn into the cell at reduced pressure. Most frequently, a multipass absorption cell based on the design of White (1942) or Herriott (1964) is employed as the sampling cell. Using base paths of 0.3 to 1.5 meters, such cells result in total absorption pathlengths ranging between 10 and 200 meters. The transmitted light is recorded with a solid state detector as the laser wavelength is repetitively scanned through absorption features of the analyte gas. Wavelength scanning is generally accomplished by applying a ramp signal (typically a sawtooth ramp) to the quiescent injection current at frequencies of 10 - 100 Hz.

Quantitative information is obtained using the Beer-Lambert Law previously discussed. However, since the diode laser source is scanned over an entire absorption feature, including the baseline off the feature, the sample gas in this case does not have to be removed to obtain the incident intensity (I_o). One uses the transmitted intensity at line center, together with measurements of pressure, pathlength, and absorption cross section, to obtain a concentration. Alternatively, one frequently integrates over the entire absorption feature using the integrated absorption cross section to obtain a concentration. This procedure, although requiring more work to accurately calibrate the frequency scale and baseline position, does not suffer from a potential systematic error due to a laser "slit function." This arises from the small but finite laser linewidth relative to the absorption feature under study. Both methods are direct absorption approaches which yield absolute concentration determinations based solely on photometric and cross section measurements, without the need for calibration standards. In fact, this approach has been used to independently determine the concentration of various calibration standards with absolute accuracies in the 1 to 10% range (see for example, Fried et al., 1990; Fried et al., 1991a). This aspect of absolute calibration sets spectroscopic techniques in general, and TDLAS specifically, apart from other measurement approaches.

In addition to direct absorption, TDL measurements are frequently carried out using the technique of harmonic detection. Most frequently, second harmonic detection is employed. In this mode, an external modulation waveform in the kilohertz frequency domain is simultaneously superimposed on the diode laser scanning current. A lock-in amplifier is used to detect the resultant frequency-modulated absorption synchronously. In most instances, the second harmonic frequency (detection at twice the modulation frequency) is chosen. Figure 6.6 illustrates the direct, first, and second harmonic absorption lineshapes:

 

Figure 6.6:   Illustration of the direct, first, and second harmonic absorption lineshapes.

As shown, second harmonic detection produces a zero baseline signal, thus eliminating the necessity of measuring small differences between two large intensities, I and I_o, as is the case for direct absorption. Further advantages of second harmonic detection over direct absorption are: (1) the elimination of a strongly sloping background often present in direct absorption; (2) reduced susceptibility to low frequency noise due to the kHz detection regime; and (3) enhanced discrimination against signals that do not have a strong wavelength dependence such as the broad absorption tails of ambient H₂O vapor.

Employing second harmonic detection, minimum detectable absorbances (ln I_o/I) of 10^-5 to 10^-6 are frequently obtained in TDLAS systems using total pathlengths around 100 m. This corresponds to minimum detectable concentrations ranging between a few parts-per-trillion (pptv) to parts-per-billion (ppbv), depending upon the absorption cross section. Molecules such as carbon disulfide (CS₂), CO₂, and carbonyl sulfide (OCS), which have very high cross sections, are examples of the former sensitivity range, while H₂Srepresents the latter. Unlike direct absorption, many instrument and experiment-dependent factors must be taken into account when deducing absolute concentrations from the measured second harmonic response. As a result, accurate quantitative analysis employing second harmonic detection is most frequently accomplished using calibration standards.
Applications

Because of the versatility, selectivity, and sensitivity of TDLAS, numerous laboratory, ground-based, aircraft, and balloon-borne studies employ this technique. Although many such examples can be cited, due to limited space, we only present here two recent applications of this technique: ground-based measurements of long-lived gases carried out at the National Center for Atmospheric Research (NCAR) and airborne measurements of carbon monoxide (CO) and methane (CH₄) carried out by researchers at NASA Langley (Sachse et al., 1991). Long-lived trace atmospheric gases such as CO₂, CH₄, OCS, and nitrous oxide (N₂O), to name a few, have received considerable attention in recent years. The ambient fluctuations of these gases contain important information about sources, sinks, and potential secular trends. In air not directly influenced by local sources, such fluctuations can be quite small, typically less than a few percent. Extracting ambient information thus requires comparable measurement precision which can be achieved by TDLAS. This, however, necessitates extreme care in both TDL system design and operation. A versatile TDL system, recently described by Fried et al. (1991b), has been developed at NCAR which addresses the major factors affecting TDL precision. This system, which is being employed in measuring the important sulfur gas OCS, can be used to measure any of the above gases with a precision around 0.1 - 0.2%. The TDL system is shown in Fig. 6.7:

 

Figure 6.7:   Schematic of optical layout and sampling system used in NCAR's TDL system.

Up to four high temperature operation diode lasers ($T \geq$ 80 K) are mounted on a temperature controlled stage housed in a liquid nitrogen dewar. The emerging infrared radiation from the selected laser, which is emitted downward, is collected and imaged into a 1.5-m base path multipass White cell. The total pathlength typically employed ranges between 90 and 150 meters and the cell pressure is around 25 Torr.

In the second application, a TDL system has been used to measure CO and CH₄ on an aircraft platform to achieve all the advantages discussed previously in addition to broad spatial coverage (Sachse et al., 1991). Although fundamentally similar to the NCAR system, the airborne spectrometer known as DACOM uses two independent laser channels to access simultaneously the absorption lines from CO and CH₄ in the 4.7 and 7.6 $\mu$m regions, respectively. These airborne measurements are extremely useful in characterizing the geographical distribution of both gases, detecting the chemical signature of biomass plumes conveyed by long range transport, identifying air mass changes and their potential origins, and in studying vertical transport. 

  
 6.3.4 Resonance and Laser-Induced Fluorescence  6.3 Measurements of Chemical Composition  6.3.2 Remote Sensing from the Ground and Aircraft 
http://www.asp.ucar.edu